From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from esa7.dell-outbound.iphmx.com ([68.232.153.96]:32067 "EHLO esa7.dell-outbound.iphmx.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1730636AbeG0T5s (ORCPT ); Fri, 27 Jul 2018 15:57:48 -0400 Cc: , , , , , , , From: To: , Subject: Re: Should a PCIe Link Down event set the PCI_DEV_DISCONNECTED bit? Date: Fri, 27 Jul 2018 18:23:30 +0000 Message-ID: <37ee3fa4e6334b5c8139a8d148fcf26a@ausx13mps321.AMER.DELL.COM> References: <47727551-86ce-040a-2516-efa47ee3a76e@gmail.com> <20180727071813.GA6128@wunner.de> <20180727170543.GA5326@wunner.de> <99604d46a7554eb38ee6c1579c53d835@ausx13mps321.AMER.DELL.COM> Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Sender: linux-pci-owner@vger.kernel.org List-ID: On 07/27/2018 01:18 PM, Sinan Kaya wrote:=0A= > On 7/27/2018 10:51 AM, Alex_Gagniuc@Dellteam.com wrote:=0A= >>> The first problem with that patch was that I hadn't fully understood=0A= >>> yet when to set PCI_DEV_DISCONNECTED and when not to set it.=0A= >> I think PCI_DEV_DISCONNECTED is a documentation issue above all else.=0A= >> The history I was given is that drivers would take a very long time to= =0A= >> tear down a device. Config space IO to an nonexistent device took a long= =0A= >> while to time out. Performance was one motivation -- and was not documen= ted.=0A= >>=0A= > =0A= > Completion timeouts are typically in the orders of 50ms. If you have a=0A= > lot of outstanding non-posted requests, it will take time for SW to=0A= > flush all requests.=0A= =0A= That's why we check pci_dev_is_disconnected() in pci/access.c=0A= But if the DISCONNECTED bit doesn't get set, it makes me wonder how =0A= solid things currently are.=0A= =0A= Alex=0A=